Evaluating the potential for workload based driving assistance systems from a psychological, technological and physiological perspective

Abstract

Advanced Driving Assistance Systems (ADAS) support the driver during highly automated and continuous driving tasks (e.g. adaptive cruise control) and provide warnings in potentially dangerous situations (e.g. forward collision warning). As ADAS do not work flawlessly, andbecause of legal issues, the driverneeds to supervise these systems in order to intervene when necessary. This means that the driver has to continually monitor information given by the systems, even though the human brain is not optimised for prolonged and monotonous control tasks, which can be interrupted by highly critical situations at any point in time. As has been shown by several validations of the Yerkes-Dodson law, the driver is vulnerable to 'out-of-the-loop' problems in low workload conditions and may not realise quickly enough that a situation is critical. Conversely, in high workload conditions the driver loses sight of the control task or needs more time to initiate an appropriate reaction in a dangerous situation. This chapter describes how a new theory has been developed, and a compensation strategy for the use of adaptive cruise control in high workload conditions has been designed and evaluated in a simulator and in on-road conditions. The technological possibility of detecting high workload conditions using physiological data has been established; the workload when the drivers' reaction time is influenced by a secondary task has been evaluated. In addition, as the usage of a system is strongly dependent on behaviour and impressions, the chapter describes how acceptance, as well as awareness of the system, has been examined.